Galectins are proteins with a characteristic carbohydrate recognition domain (CRD) (Barondes et al., 1994; Leffler et al., 2004) (FIG. 1a). This is a tightly folded β-sandwich of about 130 aa (about 15 kDa) with the two defining features 1) a β-galactose binding site (C in FIG. 1a) and 2) sufficient similarity in a sequence motif of about seven amino acids, most of which (about six residues) make up the β-galactose binding site. However, adjacent sites (A, B, D, E in FIG. 1a) are required for tight binding of natural saccharides and different preferences of these give galectins different fine specificity for natural saccharides.
The recent completion of the human, mouse and rat genome sequences reveal about 15 galectins and galectin-like proteins in one mammalian genome with slight variation between species (Leffler et al., 2004; Houzelstein et al., 2004).
Galectin subunits can contain either one or two CRDs within a single peptide chain. The first category, mono-CRDs galectins, can occur as monomers or dimers (two types) in vertebrates. The by far best studied galectins are the dimeric galectin-1, and galectin-3 that is a monomer in solution but may aggregate and become multimeric upon encounter with ligands (Leffler et al., 2004; Ahmad et al., 2004). These were the first discovered galectins and are abundant in many tissues. However, our recent phylogenetic analysis (FIG. 2) suggest that galectins with two CRDs within a peptide chain, bi-CRD galectins, appear to be more ancient and more central to the family than previously thought and that most of mammalian mono-CRD galectins may have descended from one or the other CRD of a bi-CRD galectin.
There are now over 1200 publications on galectins in PubMed, with most, as mentioned above, about galectins-1 and -3. Strong evidence suggests roles for galectins in e.g. inflammation, cancer, and development recently reviewed in a special issue (Leffler (editor), 2004b) but a unifying model of the “basic function” of galectins at the cellular-molecular level is lacking.
Galectins are synthesized as cytosolic proteins, without a signal peptide on free ribosomes. Their N-terminus is acetylated, a typical modification of cytosolic proteins, and they reside in the cytosol for a long time (not typical of secreted proteins). From there they can be targeted to the nucleus, specific cytososlic sites, or secreted (induced or constitutively) by a non-classical (non-ER-Golgi) pathway, as yet unknown, but possibly similar to the export of e.g. IL-1 (Leffler et al., 2004). They can also function in all these compartments; for galectin-3, solid evidence published in well respected journals support roles in RNA splicing in the nucleus, inhibition of apoptosis in the cytosol, and a variety of extracellular effects on cell signaling and adhesion (Patterson et al., Ochieng et al., Takenaka et al., Hsu et al. and others in Leffler (editor), 2004b). Gallectin-7 and -12 also act in the cytosol by enhancing apoptosis and regulating the cell cycle and differentiation in certain cells (Hsu and Liu in Leffler (editor), 2004b). Most galectins act also extracellularly by cross-linking glycoproteins (e.g. laminin, integrins, and IgE receptors) possibly forming supramolecular ordered arrays (Brewer et al., 2002) and may thereby modulate cell adhesion and induce intracellular signals.
The present invention relates mainly to galectin-3, but its principles may be applicable also to other galectins.
Potential therapeutic use of galectin-3 inhibitors. Galectin-3 has been implicated in diverse phenomena and, hence, inhibitors may have multiple uses. It is easy to perceive this as a lack of specificity or lack of scientific focus. Therefore, the analogy with aspirin and the cyclooxygenases (COX-I and II) is useful. The COXs produce the precursor of a wide variety of prostaglandins and, hence, are involved in a diverse array of biological mechanisms. Their inhibitors, aspirin and other NSAIDs (non-steroid anti-inflammatory drugs), also have broad and diverse effects. Despite this, these inhibitors are very useful medically, and they have several different specific utilities.
So if galectins, like COXs, are part of some basic biological regulatory mechanism (as yet unknown), they are likely to be ‘used by nature’ for different purpose in different contexts. Galectin inhibitors, like NSAIDs, are not expected to wipe out the whole system, but to tilt the balance a bit.
Inhibition of Inflammation.
A pro-inflammatory role of galectin-3 is indicated by its induction in cells at inflammatory sites, a variety of effects on immune cells (e.g. oxidative burst in neutrophils, chemotaxis in monocytes), and decrease of the inflammatory response, mainly in neutrophils and macrophages, in null mutant mice (chapters by Rabinovich et al., Sato et al., and Almkvist et al. in Leffler (editor), 2004b). Moreover, knock-out mice of Mac-2BP, a galectin-3 ligand, have increased inflammatory responses (Trahey et al., 1999). Inflammation is a protective response of the body to invading organisms and tissue injury. However, if unbalanced, frequently it is also destructive and occurs as part of the pathology in many diseases. Because of this, there is great medical interest in pharmacological modulation of inflammation. A galectin-3 inhibitor is expected to provide an important addition to the arsenal available for this.
Treatment of Septic Shock.
The idea of a possible role of galectin-3 in septic shock comes from our own studies (Almquist et al., 2001). Briefly, the argument goes as follows. It is known that septic shock involves dissemination of bacterial lipopolysaccharide into the blood stream, and that the pathological effects of this are mediated via neutrophil leukocytes (Karima et al., 1999). LPS does not activate the tissue-damaging response of the neutrophil. Instead, it primes the neutrophil, so that it is converted from unresponsive to responsive to other, presumably endogenous, activators. In septic shock, this priming happens prematurely in the blood stream. Endogenous activators could then induce the tissue damaging response in the wrong place and time. Several candidates have been proposed as these endogenous activators, including TNF-alfa. Inhibitors of these have been used in treatment schemes without much success (Karima et al., 1999). Since our own studies indicate that galectin-3 is a good candidate for being an endogenous activator of primed neutrophils (Almquist et al., 2001), galectin-3 inhibitors may be very useful in septic shock.
Treatment of Cancer.
A large number of immunohistochemical studies show changed expression of certain galectins in cancer (van den Brule et. al. and Bidon et al. in Leffler (editor), 2004b) Galectin-3 is now an established histochemical marker of thyroid cancer, and neoexpression of galectin-4 is a promising marker of early breast cancer (Huflejt and Leffler, 2004). The direct evidence for a role of. galectin-3 in cancer comes from mouse models, mainly by Raz et al, but also others (Takenaka et al. in Leffler (editor), 2004b). In paired tumor cell lines (with decreased or increased expression of galectin-3), the induction of galectin-3 gives more tumors and metastasis and suppression of galectin-3 gives less tumors and metastasis. Galectin-3 has been proposed to enhance tumor growth by being anti-apoptotic, promote angiogenesis, or to promote metastasis by affecting cell adhesion. From the above it is clear that inhibitors of galectin-3 might have valuable anti-cancer effects. Indeed, saccharides claimed but not proven to inhibit galectin-3 have been reported to have anti-cancer effects. In our own study a fragment of galectin-3 containing the CRD inhibited breast cancer in a mouse model by acting as a dominant negative inhibitor (John et al., 2003).
Also galectin-1 is frequently over-expressed in low differentiated cancer cells, and galectin-9 or its relatives galectin-4 and galectin-8 may be induced in specific cancer types (Huflejt and Leffler, 2004; Leffler (editor), 2004b). Galectin-1 induces apoptosis in activated T-cells and has a remarkable immunosuppressive effect on autoimmune disease in vivo (Rabinovich et al; and Pace et al. in Leffler (editor), 2004b. Therefore, the over-expression of these galectins in cancers might help the tumor to defend itself against the T-cell response raised by the host (Rubinstein et al., 2004). Null mutant mice for galectins-1 and -3 have been established many years ago (Poirier, 2002). These are healthy and reproduce apparently normally in animal house conditions. However recent studies have revealed subtle phenotypes in function of neutrophils and macrophages (as described above) and in bone formation for galectin-3 null mutants, and in nerve and muscle cell regeneration/differentiation for the galectin-1 null mutants (Leffler et al., 2004; Poirier, 2002; Watt in Leffler (editor), 2004b). Recently galectin-7 and galectin-9 null mutant mice have been generated and are also grossly healthy in animal house conditions, but have not yet been analysed in detail. The differences in site of expression, specificity and other properties make it unlikely that different galectins can replace each other functionally. The observations in the null mutant mice would indicate that galectins are not essential for basic life supporting functions as can be observed in normal animal house conditions. Instead they may be optimizers of normal function and/or essential in stress conditions not found in animal house conditions. The lack of strong effect in null mutant mice may make galectin inhibitors more favorable as drugs. If galectin activity contributes to pathological conditions as suggested above but less to normal conditions, then inhibition of them will have less unwanted side effects.
Known Inhibitors
Natural Ligands.
Solid phase binding assays and inhibition assays have identified a number of saccharides and glycoconjugates with the ability to bind galectins (reviewed by Leffler, 2001 and Leffler et al., 2004). All galectins bind lactose with a Kd of 0.5-1 mM. The affinity of D-galactose is 50-100 times lower. N-Acetyllactosamine and related disaccharides bind about as well as lactose, but for certain galectins, they can bind either worse or up to 10 times better. The best small saccharide ligands for galectin-3 were those carrying blood group A-determinants attached to lactose or lacNAc-residues and were found to bind up to about 50 times better than lactose. Galectin-1 shows no preference for these saccharides.
Larger saccharides of the polylactosamine type have been proposed as preferred ligands for galectins. In solution, using polylactosamine-carrying glycopeptides, there was evidence for this for galectin-3, but not galectin-1 (Leffler and Barondes, 1986). A modified plant pectin polysaccharide has been reported to bind galectin-3 (Pienta et al., 1995).
The above-described natural saccharides that have been identified as galectin-3 ligands are not suitable for use as active components in pharmaceutical compositions, because they are susceptible to acidic hydrolysis in the stomach and to enzymatic degradation. In addition, natural saccharides are hydrophilic in nature, and are not readily absorbed from the gastrointestinal tract following oral administration.
Synthetic Inhibitors.
Saccharides coupled to amino acids with anti-cancer activity were first identified as natural compounds in serum, but subsequently, synthetic analogues have been made (Glinsky et al., 1996). Among them, those with lactose or Gal coupled to the amino acid inhibit galectins, but only with about the same potency as the corresponding underivatized sugar. A chemically modified form of citrus pectin (Platt and Raz, 1992) that inhibits galectin-3 shows anti-tumor activity in vivo (Pienta et al., 1995; Nangia-Makker et al., 2002).
A divalent form of a lactosyl-amino acid had higher potency in a solid phase assay (Naidenko et al., 2000; Huflejt et al., 2001; Huflejt and Leffler, 2004) and clusters having up to four lactose moieties showed a strong multivalency effect when binding to galectin-3, but not to galectin-1 and -5 (Vrasidas et al., 2003). Cyclodextrin-based glycoclusters with seven galactose, lactose, or N-acetyllactosamine residues also showed a strong multivalency effect against galectin-3, but less so against galectins-1 and -7 (André et al., 2004). Starburst dendrimers (André et al., 1999) and glycopolymers (Pohl et al., 1999; David et al., 2004), made polyvalent in lactose-residues, have been described as galectin-3 inhibitors with marginally improved potency as compared to lactose. The aforementioned synthetic compounds that have been identified as galectin-3 ligands are not suitable for use as active components in pharmaceutical compositions, because they are hydrophilic in nature and are not readily absorbed from the gastrointestinal tract following oral administration.
Natural oligosaccharides, glycoclusters, glycodendrimers, and glycopolymers described above are too polar and too large to be absorbed and in some cases are large enough to produce immune responses in patients. Furthermore, they are susceptible to acidic hydrolysis in the stomach and to enzymatic hydrolysis. Thus, there is a need for small synthetic molecules
Thiodigalactoside is known to be a synthetic and hydrolytically stable, yet polar inhibitor, approximately as efficient as N-acetyllactosamine (Leffler and Barondes, 1986). A library of pentapeptides provided inhibitors against galectin-1 and -3, but only with low affinities, similar to that of galactose (Arnusch et al., 2004). Furthermore, peptides are not ideal agents for targeting galectins in vivo, as they are susceptible to hydrolysis and are typically polar. N-Acetyllactosamine derivatives carrying aromatic amides or substituted benzyl ethers at C-3′ have been demonstrated to be highly efficient inhibitors of galectin-3, with unprecedented IC50 values as low as 4.8 μM, which is a 20-fold improvement in comparison with the natural N-acetyllactosamine disaccharide (Sörme et al., 2002; Sörme et al., 2003b). These derivatives are less polar overall, due to the presence of the aromatic amido moieties and are thus more suitable as agents for the inhibition of galectins in vivo. However, said 3′-amido-derivatised compounds are still susceptible to hydrolytic degradation in vivo, due to the presence of a glycosidic bond in the N-acetyllactosamine disaccharide moiety and, although they are the best reported small molecule inhibitors of galectin-3, even further improved affinity is desirable.
Thus, there is still a considerable need within the art of inhibitors against galectins, in particular of galectin-3.